Using previous models to bias structural learning in the hierarchical BOA

Estimation of distribution algorithms (EDAs) are stochastic optimization techniques that explore the space of potential solutions by building and sampling explicit probabilistic models of promising candidate solutions. While the primary goal of applying EDAs is to discover the global optimum or at least its accurate approximation, besides this, any EDA provides us with a sequence of probabilistic models, which in most cases hold a great deal of information about the problem. Although using problem-specific knowledge has been shown to significantly improve performance of EDAs and other evolutionary algorithms, this readily available source of problem-specific information has been practically ignored by the EDA community. This paper takes the first step toward the use of probabilistic models obtained by EDAs to speed up the solution of similar problems in the future. More specifically, we propose two approaches to biasing model building in the hierarchical Bayesian optimization algorithm (hBOA) based on knowledge automatically learned from previous hBOA runs on similar problems. We show that the proposed methods lead to substantial speedups and argue that the methods should work well in other applications that require solving a large number of problems with similar structure.     

Human motion retrieval from hand-drawn sketch

The rapid growth of motion capture data increases the importance of motion retrieval. The majority of the existing motion retrieval approaches are based on a labor-intensive step in which the user browses and selects a desired query motion clip from the large motion clip database. In this work, a novel sketching interface for defining the query is presented. This simple approach allows users to define the required motion by sketching several motion strokes over a drawn character, which requires less effort and extends the users’ expressiveness. To support the real-time interface, a specialized encoding of the motions and the hand-drawn query is required. Here, we introduce a novel hierarchical encoding scheme based on a set of orthonormal spherical harmonic (SH) basis functions, which provides a compact representation, and avoids the CPU/processing intensive stage of temporal alignment used by previous solutions. Experimental results show that the proposed approach can well retrieve the motions, and is capable of retrieve logically and numerically similar motions, which is superior to previous approaches. The user study shows that the proposed system can be a useful tool to input motion query if the users are familiar with it. Finally, an application of generating a 3D animation from a hand-drawn comics strip is demonstrated.     

Mathematical modeling of the heart using magnetic resonance imaging

A hybrid three-dimensional solid mathematical model of cardiac ventricular geometry developed using magnetic resonance (MR) images of an in vivo canine heart is discussed. The modeling techniques were validated using MR images of an ex vivo heart and direct measurements of cardiac geometry and mass properties. A spin-echo MR sequence with in-plane resolution of 1.0 mm was used to image the canine heart in eleven short-axis planes at contiguous 5-mm intervals. Contour points on the epicardial, left ventricle (LV), and right ventricle (RV) boundaries were selected manually at each slice level. A boundary representation geometric model was constructed by fitting third-order nonuniform rational B-spline surfaces through each set of surface points. Compared to the anatomic specimen (AS), volume errors of the ex vivo model were 0.3, 1.5, and 5.8% for the LV cavity, RV cavity, and total enclosed volumes, respectively. Comparison of cross-sectional areas of the AS and the model at ten levels demonstrated mean model errors of 4.1, 2.5, and 2.9% for the LV, RV, and epicardial boundaries, respectively     

多层水平气流带式干燥机的数学模型

根据干燥传质传热学原理和干燥动力学特性，建立了水平气流带式干燥机干燥过程的数学模型。该模型可以预测干燥机合理的通风道高度、带翻转最佳时间和热风风速等。为优化多层带式干燥机的结构或设计新的水平气流带式干燥机提供了理论依据。     

Three-channel solar radiometer for the determination of atmospheric columnar water vapor

The design of a three-channel solar radiometer used to determine total columnar atmospheric water-vapor amounts is presented. The main channel is located in the 0.94-μm water-vapor band, and two other channels are located in adjacent nonabsorption regions of the solar spectrum and are used to remove scattering effects from the main channel. Water-vapor transmittance is determined by means of a modified Langley approach, and these transmittances are converted to columnar water vapor by means of a band model developed at the University of Arizona. Several cases are presented in which columnar water-vapor amounts are determined through the use of the instrument and method described here. These results are compared with sounding-balloon results. Tests of the method indicate that columnar water vapor may be retrieved with an uncertainty of less than 10%.     

J. H. Kulick, J. M. Jarem, R. G. Lindquist, S. T. Kowel, and M. W. Friends are with the Department of Electrical and Computer Engineering, University of Alabama in Huntsville, Huntsville, Alabama 35899

The development and modeling of a liquid-crystal phase grating for real-time diffractive three-dimensional displays are discussed. The system being developed, which is called the ICVision system, utilizes a number of ideas that will result in a rugged, low-power three-dimensional display offering both vertical and horizontal parallax and eventually full color. Fringing fields created between interdigitated electrodes formed on top of VLSI die will induce a diffraction pattern in a thin layer of liquid crystal that will cover the die. A detailed electrostatic and diffraction analysis of liquid-crystal phase-grating regions that will make up the final display is given here. The electrostatic analysis is developed by use of the method of moments. The diffraction analysis is developed by use of rigorous coupled-wave diffraction theory. The numerical results obtrained from the mathematical model are compared with experimental diffraction results from preliminary LCD cells that have been assembled as prototype ICVision devices.     

Liquid crystal-on-silicon implementation of the partial pixel three-dimensional display architecture

We report the implementation of a liquid crystal-on-silicon, three-dimensional (3-D) diffractive display based on the partial pixel architecture. The display generates multiple stereoscopic images that are perceived as a static 3-D scene with one-dimensional motion parallax in a manner that is functionally equivalent to a holographic stereogram. The images are created with diffraction gratings formed in a thin liquid crystal layer by fringing electric fields from transparent indium tin oxide interdigitated electrodes. The electrodes are controlled by an external drive signal that permits the 3-D scene to be turned on and off. The display has a contrast ratio of 5.8, which is limited principally by optical scatter caused by extraneous fringing fields. These scatter sources can be readily eliminated. The display reported herein is the first step toward a real-time partial pixel architecture display in which large numbers of dynamic gratings are independently controlled by underlying silicon drive circuitry.